GitHub adds team-level Copilot usage metrics through a new API join workflow

GitHub now lets admins build team-level Copilot usage metrics by joining a new user-teams report with the existing per-user usage report. The release adds NDJSON download endpoints for enterprise and organization scopes, so engineering leaders can analyze adoption, active users, completions, chat activity, and model or IDE breakdowns by team. The important catch is that this ships as an API workflow, not as a built-in dashboard.

Key takeaways

• GitHub added a new user-teams report for Copilot usage metrics.
• Two new REST endpoints return signed NDJSON download URLs for one-day team membership reports.
• Team-level analysis requires joining the user-teams report with the per-user usage report on user_id and day.
• GitHub says there is no dashboard surface for this release; the feature is API-only.
• Teams with fewer than five Copilot-seated users are excluded from the user-teams report.

Why it matters

This matters more for engineering managers and platform teams than for individual developers. If your company is paying for Copilot at org or enterprise scale, total adoption numbers are rarely enough. Team-level reporting gives you a practical way to spot which teams actually use Copilot, which ones only have licenses assigned, and where rollout or enablement is underperforming.

It also helps when usage-based billing, internal ROI reviews, or support for model policy decisions start showing up in quarterly planning. You can separate “Copilot is popular somewhere” from “this backend team, support engineering group, or mobile org is using it consistently.”

What to verify before you act

First, confirm permissions and policy settings. GitHub says these metrics are limited to enterprise administrators, organization owners, billing managers, or specific enterprise custom roles, and the relevant Copilot metrics policy has to be enabled.

Second, validate how you want to count multi-team users. GitHub notes that users who belong to multiple teams will appear in each team aggregate, so summed team totals will not reproduce organization-wide totals.

Third, check your small-team blind spots. GitHub excludes teams with fewer than five Copilot-seated users from the user-teams report, which means some specialist or pilot teams may vanish from this specific view.

Practical LinkLoot angle

The fastest useful workflow is simple: schedule daily pulls of the user-teams and per-user usage reports, join them in your warehouse or notebook layer, then publish a lightweight internal scorecard that tracks adoption, chat usage, completions, and model mix by team.

That gives you enough signal to answer rollout questions without waiting for a future dashboard. If you are building internal automations around developer tooling, LinkLoot’s workflow guide is a good fit here: /guides/ai-workflow-automation.